1844. All-Cause Readmission Following Treatment with Linezolid for MRSA-ABSSSI: A Propensity Score Analysis
Session: Poster Abstract Session: Treatment of HAIs/Antimicrobial Resistant Infections
Saturday, October 10, 2015
Room: Poster Hall
Background: MRSA-associated acute bacterial skin and skin structure infections (MRSA ABSSSI)are a major cause of patient morbidity and health resource utilization. The rising prevalence of vancomycin-intermediate MRSA and the emergence of vancomycin-resistant MRSA increase the need to evaluate non-vancomycin treatments for MRSA infection, including alternatives such as linezolid.

Methods: Laboratory and billing records were abstracted for adult patients admitted to the Detroit Medical Center and Henry Ford Health System (Detroit, MI) from 2006 to 2012. All inpatients receiving linezolid were identified and randomly matched 1:3 by health system and year to patients receiving another antibiotic (vancomycin, clindamycin, daptomycin, trim-sulfa, ceftaroline, doxycycline).  Clinical and demographic variables were abstracted from medical records. Readmissions were captured at any included hospital system. A propensity score was calculated to predict treatment assignment using stepwise model selection with a threshold of 0.2. Clinical outcomes were compared between groups using logistic regression with inverse probability weighting. 

Results: 164 hospitalized patients with MRSA-ABSSSI were treated with linezolid over the six year period and were matched to 502 comparison patients. The multivariate propensity model of differences between treatment groups included hospital transfer, severe sepsis, history of previous hospitalization, current ICU admission, renal failure, bacteremia and stage 4 ulcer. The all-cause readmission rate was 25% (41/164) in the linezolid group and 20% (100/505) in the comparison group. Following propensity weighting, all-cause 30-day readmission was not statistically different between groups (OR: 1.2 (95% C.I. 0.8, 2.0); p=0.38). This estimate was significantly modified between those with diabetes (OR: 0.6 (95% C.I. 0.3, 1.3); p=0.20) versus those without (OR: 1.8 (0.98, 3.4); 0.06) (p=0.03 for interaction).

Conclusion: Our analysis did not identify a difference between inpatients with MRSA-ABSSSI treated with linezolid versus other antibiotics with regards to 30-day all-cause readmission. However, we did find a significant interaction by diabetes status. The potential for an increased treatment benefit with linezolid among patients with diabetes warrants further investigation.

Emily T. Martin, MPH, PhD, Department of Epidemiology, University of Michigan School of Public Health, Ann Arbor, MI, Richard Evans, BS, University of Michigan School of Public Health, Ann Arbor, MI, Matthew Compton, ., Pharmacy Practice, Wayne State University, Detroit, MI, Jason Pogue, PharmD, Detroit Medical Center/Wayne State University, Detroit, MI, Paul Lephart, PhD, Detroit Medical Center University Laboratories, Detroit, MI, Michael J Rybak, PharmD, MPH, Wayne State University, Detroit, MI, Keith Kaye, MD, MPH, FIDSA, FSHEA, Detroit Medical Center (DMC) / Wayne State University, Detroit, MI and Susan L Davis, PharmD, Henry Ford Hospital, Detroit, MI; Wayne State University College of Pharmacy, Detroit, MI


E. T. Martin, Pfizer: Grant Investigator , Grant recipient

R. Evans, None

M. Compton, None

J. Pogue, CUBIST: Consultant and Speaker's Bureau , Consulting fee , Research grant and Speaker honorarium
Actavis: Speaker's Bureau , Speaker honorarium
theravance: Consultant and Speaker's Bureau , Consulting fee and Speaker honorarium

P. Lephart, None

M. J. Rybak, None

K. Kaye, Detroit Medical Center and Wayne State University: Consultant and Grant Investigator , Consulting fee and Grant recipient

S. L. Davis, None

Findings in the abstracts are embargoed until 12:01 a.m. PDT, Wednesday Oct. 7th with the exception of research findings presented at the IDWeek press conferences.